Improvements in the Diffraction Process.
(Stated Meeting, held Wednesday, March 2z, z9o6.) I m p r o v e m e n t s in the Diffraction Process of Color
Photography. BY HERBERTE. IVES. The diffraction process of color photography, invented by Prof. R. W. Wood, of Johns Hopkins University, in I899, is an application of the well-known three-color method of reproducing colors by photography. This method depends primarily upon the observations of Young, Helmholz and Clerk Maxwell, that all the colors of the solar spectrum may be counterfeited to the eye by mixtures of three narrow bands of color from the spectrum, these colors are red, near the Fraunhofer line C; green, near E, and blue, near F. For instance, ied and green mix to give the eye a sensation of yellow indistinguishable from the true yellow of the spectrum; red and blue mix to give purple; and the three colors acting together produce a white whose difference from ordinary white light can be detected only by analysis with a spectroscope. What applies to spectrum colors applies equally well to the varied hues of nature. The coloring of such an object as a basket of fruit can also be duplicated to the eye by mixtures of the three primary colors. The tint o.f an apple, by a large proportion of red, less of green and blue; of a lemon, by nearly equal parts of red and green; of grapes, by a large proportion of blue. The three-color process can be reduced to two problems; first, the production of three photographic negatives, each of which shall be an exact record of the amount of one of the primary colors requisite to mix with the others and counterfeit to the eye the color of the object photographed; second, some means of furnishing each record with its appropriate color and combining it with the others. The solution of the first problem has been arrived at from experimental quantitative determinations of the mixing proportions of the primaries to produce the other colors. From these determinations three-color screens can be prepared,
440 Ires: [J. F. I., which, when used with suitable photographic plates, will yield three (black and white) negatives, each having the desired dest,.'tution of light and shade to form a record of one p r i m a r y color. The negatives thus obtained are the basis of all threecolor reproduction methods. Numerous means have been suggested and tried for c o m bining the three-color records with their corresponding colors. T h e y may be placed in a triple lantern, each illuminated with its proper colored light and projected, superposed, upon a screen. The superposition may be effected by a system of mirrors, as in the K r o m s k o p ; by the use o f three thin transparent films properly colored; by triple printing on paper, after the m a n n e r of much of the present-day magazine illustration. A process which must be noted somewhat in detail because of its direct bearing on the recent development in diffraction color p h o t o g r a p h y is the so-called Joly process*. Combination of the colors is effected in this by breaking up the threecolor records into narrow lines, arranged in succession, a line of the red record, a line of the green, a line of the blue, and so on repeating across the picture. This triple record, whose lines should be close enough t o g e t h e r to be indistinguishable by the eye, is mounted over a triple ruled color-screen,--a line of red pigment, a line of green, a line of blue, similarly spaced to t h e lines of the picture. The result, if the lines are fine e n o u g h - - a condition never yet attained in the actual working of the process,--is that the eye blends the lines to form a structureless color picture in the form of a transparenc.y. The diffraction process, which is the subject of this paper, departs widely from the other methods. Its d i s t i n g u i s h i n g feature is that for the production of the primany colors to view the records use is made of the diffraction grating, that is, of a transparent polished surface, usually of glass, ruled with fine parallel straight lines, several thousand to the inch. It is the property of a diffraction grating that if a bright line or point of light is viewed through it, not onlywill the light source be seen. but spread out ~:o either side will be a series of spectra, those nearest the source being called spectra of the first order, the next, of the second order, etc. If the number of lines to t h e * First published, as a matter of fact, by Louis Ducos du Hauron in i859.
Improvements in the Diffraction Process.
inch on the grating be increased the spectra are thrown farther from the central image, and vice versa. The power of a grating to produce color is taken advantage of in the following way : Suppose we have a convex lens forming an image on a screen of a bright source of light, such as a gas flame. If the eye is placed where the image is formed the lens is seen uniformly and brilliantly illuminated. S u p p o s e now a diffraction grating is placed over the lens. In addition to the image formed as before there will be produced a series of spectra. If the eye is placed in one of these the lens will, as before, appear illuminated, not, however, by light of the color of the source, but by the c o l o r o f light striking the eye. If now we can make one of our color records in the form of a diffraction grating of varying strength to correspond to the desired differences in the amount of the primary color, and
place it over a lens, points can be found in the lateral sPectra in which the lens (and the grating in coincidence with it) will appear as a colored picture. Further, since, as we have seen above, the distance of the spectra from the central image depends upon the fineness of the grating spacing, it is a simple m a t t e r to choose three gratings, one of which will send red to a chosen point, the second green, the third blue. H e n c e if we can make the three primary color records in the form of three diffraction gratings of three properly chosen spacings, each may be seen in its proper color by placing the eye in one of the diffraction spectra formed as above described. In Fig. I we have represented the conditions for viewing diffraction color pictures. A is a source of light, B a convex lens, in front of which are three gratings G. On the screen D fall the central image I, and three spectra (only the first order spectra on one side are represented) so placed that the red of one,
[J. F. I.,
the blue of another, and the green of the third are superposed on the slit S, at which the eye is placed. In Fig. 2 we have represented diagramatically a diffraction color picture of a red flower with green leaves on a blue ground. ~t'he coarse spacing of the lines in the flower represents a grating to send red light to the eye, say 24o0 lines to the inch, the medium spacing of the leaf one to send green to the eye, say 3000 lines to the inch, the fine spacing of the background one ~o send blue, say 3600 lines to the inch. Mixed colors would be given by two or three gratings acting together. To produce such gratings by photographic means the bichromated gelatine process, which lends itself well to the copying of minute structures, was used. In contact with a surface sensitized in this way was placed a glass grating; the image of the corresponding color record was then projected upon the surface
for a sufficient time to give a full exposure. The grating was removed, another substituted and exposed under its corresponding color record, and so with the third. In this way all three grating pictures were printed, one on top of another,* forming a picture which by diffused light was transparent and quite invisible, showing its color only when viewed with the proper combination of lens and bright source of light. F r o m the pictures made in this way copies could be made by simple contact printing on bichromated gelatine. Since a direct copy of a grating is still a grating, i. e., a series of lines, the process is a positive one and copies are not reversed in light and shade as in making copies of ordinary photographs. It is obvious that quite apart from its scientific interest the * I n p r a c t i c e it w a s f o u n d i m p o s s i b l e to g e t t h r e e i m p r e s s i o n s o n o n e gelatine surface and so two were made on one surface and the third on a n o t h e r , t h e t w o s u r f a c e s b e i n g a f t e r w a r d s p l a c e d in c o n t a c t .
Improvements in the Diffraction Process.
diffraction process promises very real advantages. F o r instance, the colors used are beyond question pure spectrum colors, and so there is no need to depend on dyes or colored glasses; also the ease and cheapness with which copies can be made places it in a class by itself a m o n g three-color processes. ,So perfect indeed did the process seem theoretically when first published that there was every reason to exp.ect results fully comparable with the best of other methods. This early promise was not fulfilled. A few pictures were .obtained, interesting as scientific curiosities only. N o dependd e n c e could be placed in the results; some colors reproduced well, others did not; occasionally a good picture would be made, but the same procedure applied to another subject brought no success. Six years after its publication the process ibad made no progress and seemed fated to rank as a failure. Last summer, through the courtesy of Prof. W o o d , the writer was loaned a number of diffraction gratings, ruled on the Rowland dividing engines at Johns Hopkins University. Experiments with these revealed a fundamental defect on the .above-described mode of making diffraction pictures. By finding means to overcome this defect results have been obtained .of a remarkable degree of perfection. The defect referred to is that the three gratings, in order to g e t their joint effect, were superposed, being, as we have seen, printed one on the other. In so doing the assumption was made that the effect of superposing gratings was to add their separate effects. As a matter of fact, additional, disturbing effects are introduced, partly due to the inability of the gelatine surface to take several grating impressions without mutual blotting out, and partly--chiefly, in f a c t - - t o the forming of a n e w c o m p o u n d grating. That is, if two gratings of different spacings are superposed, the two spacing periodically get in and out of step with each other, and this new periodic structure forms itself a diffraction grating. The new grating then forms its own series of spectra, which subtract light from the •original ones. Therefore when the t w o gratings are superposed, the eye, instead of receiving a double quantity of light receives much less than the double quantity. Even more serious than this loss of light is the fact that the new spectra due to the two gratings t o g e t h e r frequently fall in such a position as
[J. F. I.~
to introduce false colors. This is well illustrated by taking two, gratings of different spacing and placing them on one another at right angles. Two sets of spectra will be formed, one by each grating, and parallel to it, and, in addition, a number of diagonally disposed spectra. As the gratings are turned into. the same straight line all the spectra turn, and the additional_ diagonally-placed spectra take up positions between the spectra formed by the original gratings. Consequently, while the eye may receive red from one grating and blue from. another, one of the spectra due to the two together may send some o t h e r color, such as green. This case actually occurred frequently, a pink rose reproducing as green, and red and blue color disc~ superposing to give green instead of purple. These observations made clear the necessity for some method of obtaining the effects of the three gratings o t h e r than by superposition. It was at once seen that this could be ,~.ccomplished by a procedure similar to the Joly process, e,amely, by having the grating elements in narrow juxtaposed strips. Some experiments had already been made by Prof. W o o d with Joly pictures, not, however, with the specific purpose above mentioned, but rather to illustrate to possibility of making such pictures with very much finer color lines than it i~ possible to do by ruling alternating colored pigment lines for the observing screen. The mode of procedure involved' laboriously ruling a special grating consisting of several lines. of one spacing, followed by several of another, and then several of the third, repeating all the way across the plate, The widtl~ of each strip of lines was made to correspond to the width of arc element of the Joly picture. F r o m this grating a print wa~ made on the special-line picture, which had been previously flowed with gelatine. This in turn was used to print gelatine copies. A practical disadvantage of this plan, aside from the use of the special grating, is that one is restricted to the use of original Joly pictures of a certain definite spacing of line, determined by the limitations of the process employed in their p r o d u e tion. A much more serious defect arises, however, in this way: T h e "Joly lines" if made, as they should be made, several hundred to the inch, themselves form a diffraction grating, which, as it is parallel to the three principal gratings, forms spectra
Improvements in the Di#raction Process.
superposed on those depended on to reproduce the colors of the object. This is quite as serious a defect as that arising from superposed gratings, and is sufficient to condemn the procedure. F r o m a consideration of these various difficulties it followed that some means of breaking the picture up into lines was imperative, and that that means should not involve the use of a special grating, nor of special Joly original pictures, difficult to obtain, and, most important of all, the narrow color strips or j o l y lines must be arranged in some way so as not to give disturbing grating effects. All of these ends were achieved by the following procedure:
In F i g . 3, which represents the method of making the imp r o v e d diffraction pictures, A is the bichromated gelatine plate, rigidly fixed in position; B is a glass diffraction grating; C is a line screen, ruled with at least two hundred lines to the inch, with the opaque lines twice the width of the transparent ;* D a lens, and E a positive color-record to be copied. The latter is an ordinary three-color positive containing no lines or s t r u c t u r e , f and the grating is an ordinary continuously-ruled one. With say the red record at E and the corresponding grating at 13, an exposure is made, resulting in a series of narrow ~rips. A second positive is then placed at E, the corresponding grating at B and the ruling C moved the midth of a transp a r e n t portion. A second exposure is then made, the opaque * The opaque line screens were ruled by Mr. Max Levy, to whom, for his interest and generous assistance, the writer is greatly indebted. Posi~ves from negatives made for the Kromskop were used.
[J. F. I.,
lines shielding the previously exposed surface, and a similar t r e a t m e n t given to the third positive. There results finally a picture made up of alternating strips of three different gratings. To eliminate the grating effects of the narrow strips of gratings considered as lines, the device is used of making the strips (Joly lines) run at right angles to the diffraction grating lines, so that the spectra produced by them are thrown off in another direction and do not enter the eye. Although the device is simple it is of extreme importance, and its adoption is rendered possible only by the plan described for m a k i n g the pictures. The difficulties in the way of ruling a special grating with the three gratings disposed in a similar manner are practically in-
~ ~-.~---=--~l 0 '*"-3000l;nt~ . to . inch . "-
superable. It is obvious that the strips of grating can be made as narrow as desired, easily narrow enough to be indistinguishable as such by the eye. Fig. 4 gives an idea of the appearance of the finished picture under the microscope. The short, fine lines are the diffraction grating lines furnishing the three primary colors; 24oo to the inch for the red, 3ooo for the green, and 36oo for the blue. The broad strips at right angles to the grating lines constitute the "Joly lines," of which there should be at least 2oo groups of three to the inch. W h e n viewed with a lens and bright source of light the pictures made in this way are entirely free' from the formerly-obtained defects. The colors are pure and brilliant, and, unlike ordinary Joly pictures, the color lines are too fine to be visible.
June, I9o6.] Improvements in the Diffraction Process.
The results indeed approach those obtained with the Kromskop. As a further modification of the original method the writer has found it possible to dispense with three gratings and obtain the colors with a single grating spacing properly used, To do this the source of light must be a rather long slit. Viewed through a grating the slit of course gives long spectra parallel to its length. If now the grating be rotated about, the perpendicular dropped from it to the slit, the spectra move in toward the slit. [The accompanying shift parallel to the length of the slit is compensated for by the slit being long.] So, by suitable rotation any desired spectrum color may be obtained at a chosen point. Starting with a grating of 36oo lines to the inch to give the blue when parallel to the slit, a rota-
tion of about 21½ degrees will give the green, of 42 degrees the red. In the absence of suitable dividing engines to rule three properly-proportioned gratings this affords an exact and easy method of securing the three colors. It has the fourth advantage, that in printing copies such difficulties as securing perfect printing contact will affect all three colors alike, which is not the case with gratings of different degrees of fineness. Fig. 5 shows a portion of a picture made in this way with one grating spacing.* *After w o r k i n g out this idea the writer learned that some years ago Mr. Thorp, of Manchester, suggested the use of a single g r a t i n g spacing to secure all three colors. Mr. T h o r p ' s plan, however, was to use t h r e e sources of light and merely rotate the gratings until they "found" the source and each cleared the source b e l o n g i n g to the other two. H e found a rotation of ten degrees convenient. As far as the writer knows this is the first publication of a plan to secure any desired color by rotation t h r o u g h a definite angle to be calculated from the wave length.
[J. F. I.,
With these improvements probably the last word has been said on the diffraction pictures themselves. A very important improvement in the means for observing them, due to the writer's father, Mr. Frederic E. Ives, must be described. The lens and bright light used by Prof. W o o d do not form at all a conveinent arrangement, nor is it desirable to use artificial light. A convenient apparatus, easily set up, not liable to get out of order, and suitable for daylight use became desirable as soon as the pictures are perfected. The instrument about to be described was devised a few hours after the first pictures were obtained, and admirably fulfills its purpose. The greatest difficulty attending the use of daylight is that of getting sufficient light,--the illumination of the sky, toward which an instrument would naturally b~, pointed, is far from, intense e n o u g h . This will be appreciated when it is r e m e m b e r e d that only a very small portion of the original light is diffracted, perhaps ten per cent. at most. This difficulty has been overcome in a novel manner. Instead of depending on a single slit, as the narrow source of light, a series of slits is used, each furnishing one spectrum. In this way, with four slits, two first order and two second order spectra are utilized, yielding probably three times the light obtainable from a single slit. Fig. 6 gives the instrument in section. A, B, C, D are the four slits; M a mirror; L1 and L~ lenses; P the diffraction picture; and S the slit t h r o u g h which the picture is observed. The lenses of course form an image of each slit at A1,B1,C1,D 1 ; from each of these images, however, a certain a m o u n t of light is diffracted by the pictureP; from B and C first order spectra fall on S, from A and D, second order. The use of second as well as first order spectra is a distinct advantage in that, as gratings never give a perfectly uniform distribution of light and color, certain desirable qualities of the picture are found in one ,order and not in the other, While if both orders are used the resultant evening up of qualities produces particularly satisfact o r y results. By disposing the grating lines in a horizontal direction and using horizontal slits as sources the pictures may be viewed by both eyes, a desirable condition for convenience and comfort. As an instrument the "Diffraction Chromoscope" is simplicity itself. It is, in fact, used much as the old stereoscope.
June, 19o6.1 Improveme~lts i;z the Di#ractio;~ Proccss.
There are no adjustments; to use, it is merely placed before a window or Welsbach light and the pictures.dropped to place. O n looking into the eye slit before the introduction of the picture nothing is seen, the inside being perfectly black. The pictures themselves are transparent, colorless, and appear as plain pieces of glass under ordinary conditions of illumination. O n placing them in the instrument the colors immediately flash out, a transformation which seems almost magical, affording a scientific demonstration of rare beauty.
Aside from the obvious use of the apparatus for scientific purposes it is expected that its simplicity and the perfection of the results wiJl ultimately lead to many important uses. N o w that the long standing obstacles in the way of success have been removed the process should develop rapidly. Such further steps as application to lantern projection and means for making the pictures directly in the camera are under consideration.